On 2021-03-10 at 01:30, Arto Jantunen wrote: > Indeed the package was rejected after two months in the queue, due to > things missing from the copyright file: > >> +----------------------+ >> | REJECT reasoning | >> +----------------------+ >> >> examples/completion seems to be copyright Etienne PIERRE and there does not >> seem to be reason that they too have relinquished copyright. >> >> moosic/server/xmlrpc_registry.py has a different license. >> >> +----------------------+ >> | N.B. | >> +----------------------+ >> >> This review may not be exhaustive. Please check your source package >> against your d/copyright and the ftpmaster REJECT-FAQ, throughly, >> before uploading to NEW again.
Neither of these things is new; they were true of the last version prior to the removal, and possibly of some versions prior to that as well. That makes this a bit aggravating. Still, I suppose that just means they slipped through the cracks of the less-stringent copyright review that's applied to packages already in the archive, rather than that they shouldn't need to be addressed... > I'll try to find the time to go through the source and update this, > unless you beat me to it of course. :) For moosic/server/xmlrpc_registry.py, I think we just need to document the license in debian/copyright. I don't have a local copy of the debian/ directory for this, and have no experience with updating such, so although I'd kind of like to *get* that experience I think it'd probably be best if you cover that part. For examples/completion, it's not clear whether or not documenting the copyright statement would be enough. No specific license is stated for that file, so it's not clear what license Etienne PIERRE (whom I infer to be its original author, prior to later changes introduced by Daniel Pearson) would have intended for it. It's possible that the lack of license statement for that file indicates that it's covered by the Unlicense, like the rest of the codebase; however, since the rest of the codebase except for xmlrpc_registry.py seems to have all come from Daniel Pearson, I wouldn't necessarily want to assume that. Based on the copyright statements in that file and their dates, my guess is that moosic originally didn't provide completion, that a user contributed that functionality, and that the maintainer later iterated on the base of that contribution. If that's correct, then the license for this file would probably be whatever license moosic was under at the time of its introduction; we could probably find that out by examining the versions of the package from prior to the relicensing, which I think took place in 2011. We might need to contact Etienne PIERRE, to ask about what license was intended here, and whether the change to the Unlicense would be OK. Fortunately, an E-mail address is provided; unfortunately, it's 18 years old, so could well be out of date. As a third possibility, we could argue that this completion file is simple enough to not be meaningfully copyrightable, but I wouldn't want to try to make and back up that argument in the face of FTP-team copyright review. As a fourth possibility, we could investigate whether this file is even needed anymore. Nothing else in the source seems to reference it, except for setup.py (apparently as part of building the documentation?); the ChangeLog references tab-completion once, a 'debian/completion' file which is no longer part of the upstream source (it's even possible I'm the one who killed it, since I don't think debian/ should be maintained upstream), and tab-completion functionality of a long-dead specialized shell called 'moosh'. It may be notable that the dates of those references are all around 2003, which is the year of Etienne PIERRE's copyright on examples/completion. Given that I don't use completion in a moosic context (or programmable tab-completion at all in my shell), I'm not sure how to effectively test whether this is still needed or not. Suggestions are welcome. (That said, a simple cmp shows that it's identical to /etc/bash_completion.d/moosic from the existing 1.5.6-1 package, so apparently we have been shipping this in Debian.) I've done a cursory glance into every file in the source (in the form I last saw it on GitHub; I didn't bother to re-pull in case someone changed it there under my feet), and other than the above, everything else either states no license or explicitly states the Unlicense. The files that state no license seem to mostly be Python boilerplate, build-system glue (the Makefile, setup.py), or in the doc/ subdirectory; a few of them (instances of __init__.py) are actually empty. I think it's fair to assume that they're all Unlicense'd, and written by Daniel Pearson himself. We could always ask him (including questions about the two files noted above, in relation to the relicensing), but since I already inquired and he indicated that he no longer has interest in moosic, I don't want to impose on him unnecessarily. > In any case we have missed the freeze by a mile, so we have a couple > of years to get this done before the next round. I'd ideally prefer to at least have it in unstable (or at worst experimental) before the next stable release, but yeah, no rush to beat the freeze proper at this point. -- The Wanderer The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw
signature.asc
Description: OpenPGP digital signature
